14 DAY TRIAL // Google has been customizing search results for particular users for years now, gradually introducing new rules and filters. At this point, one in five Google searches will be specific to the user doing the search and will differ from the vanilla search results ranking. This may seem trivial but Google spends a lot of time and resources trying to make the results more relevant to each individual user and, sometimes controversially, it uses quite a bit of personal data for this.
The company employs just three main data points to customize the search results for each individual user, location, web history and online contacts (social search). "As it stands today," Google software engineer Bryan Horling told The Register, "between these three techniques, just about every user who's engaging with Google search today is affected."
These customizations refer to users in the same countries and continents as Google has been serving different results based on that for several years. But the latest developments go beyond that and cater to individual users. Google has a good reason for this, different people have different expectations and different needs, so one raking order that is relevant for one may not be for someone else. In fact, it has recently enabled personalized results even for users who aren't logged into an account to help with the relevancy of the results in those cases as well.
However, Google says that most of the times, the differences are very small and subtle involving just a small reshuffle of the regular results. "When these techniques fire, the changes tend to be relatively minor," Horling stated. "We're moving a few results. We might be moving a few down. We're generally not changing the entire character of the page."
He says that Google is reluctant to do big changes for at least two major reasons. One is that, based on the information it has gathered, the company already knows what a certain user is looking for so, in cases where a search term may be ambiguous, the search engine has a pretty good idea what meaning the user is referring too. The second reason is that Google's algorithms aren't always right and making major changes based on their predictions may have pretty big implications.